Dictation has always had an embarrassing gap between what you say and what you actually mean to send. Every “um,” every false start, every half-formed thought — it all lands in the text box raw. Nothing thinks it has a fix.
Essential Voice, rolling out now on the Phone (3), strips filler words before the text reaches you. Not after — during. The idea is you talk, it transcribes, and what comes out the other end is something you’d actually want to send rather than a transcript that makes you sound like you’re thinking out loud in a business meeting.
It handles over 100 languages and picks up regional variants automatically, so British English and American English aren’t getting confused. Translation runs alongside transcription if you need it.
There’s a shortcut system worth knowing about. You can train it — tell it that whenever you say “nothing os” you mean “Nothing OS,” or link a restaurant name to its full address. Small thing, but anyone who dictates frequently knows how much time gets eaten fixing proper nouns and recurring spellings.
On privacy: Nothing says it only activates on demand, never sitting in the background waiting. Audio gets encrypted and sent to their servers for processing, then the text comes back to your device. Nothing claims they don’t hold onto the audio or the generated text after that.
Phone (3) owners get it today. The Phone (4a) Pro follows later this month, with the standard Phone (4a) coming in early May. It plugs into the built-in keyboard and the Essential Key.
